| 07:56 | <annevk> | philipj: perhaps Firefox throws because 2^32-1 is out of range/index? |
| 07:56 | annevk | is not super familiar with this API |
| 08:01 | <philipj> | annevk: yeah, it would still throw, just for a different reason |
| 08:03 | <annevk> | philipj: indeed, seems like the IDL can be safely changed, but then you need to decide what to do with the value... |
| 08:03 | <philipj> | yep |
| 08:03 | <annevk> | well, "safely", you will get different results |
| 09:38 | <annevk> | "[T]he reviewer for my registration of application/xhtml+xml in 2014 objected because I was not referring to XHTML 2.0." Thank you darobin, that's comedy gold |
| 09:39 | <darobin> | annevk: :) |
| 09:39 | <darobin> | yeah, MikeSmith and I laughed about that for a while |
| 09:39 | <darobin> | shame the stuff is not public |
| 09:39 | <darobin> | PLH had to stop me from basically strangling them |
| 10:26 | <annevk> | SimonSapin: https://www.w3.org/Bugs/Public/show_bug.cgi?id=26341 is so true |
| 10:26 | <annevk> | SimonSapin: and over a year old now, I guess I should try to come up with a model that sucks less |
| 10:30 | <aleray> | Hi, I'm trying to implement a filter in html5lib python to avoid word widows and orphans by replacing the first and last spaces of, let say, paragraphs with non-breaking spaces. it seems more complicate than I initially though because it should handle cases where the paragraphs have sub elements like `em` or `bold`. Any idea how to implement this? |
| 10:50 | <annevk> | In IE11, if you parse file:#x or file:?x, pathname returns U+FFFD |
| 10:50 | <annevk> | What is even happening? |
| 10:59 | <annevk> | Hmm, in IE11 on Win10 I can't connect to google.com due to some certificate error... |
| 11:37 | <annevk> | So the problem was that the time was not accurate inside VirtualBox... |
| 11:37 | <annevk> | And the time synchronization feature of Windows was broken |
| 11:38 | <annevk> | Which makes me wonder how you'd run time synchronization over HTTPS when you don't know the time yet... |
| 13:08 | <SimonSapin> | MikeSmith: I’m getting email from buzilla⊙jwo (A "g" got dropped, making email filters not match anymore. I’ve updated my filter, but maybe this also affects other people.) |
| 14:33 | <wanderview> | annevk: is there something that wants to run time synchronization over https? |
| 14:33 | <annevk> | wanderview: an OS should |
| 14:33 | <annevk> | wanderview: otherwise a network attacker can invalidate HSTS entries |
| 14:34 | <wanderview> | annevk: most OS implementations will not slew the clock by huge values from ntp updates |
| 14:34 | <wanderview> | if its outside a reasonable range it just ignored the input |
| 14:34 | <annevk> | wanderview: after I learned about this attack I think there's basically nothing you want to do without HTTPS |
| 14:35 | <annevk> | wanderview: hmm, after I found a server Windows just changed the clock by about a year |
| 14:35 | <wanderview> | annevk: thats a windows bug then :-) |
| 14:35 | <wanderview> | IMO |
| 14:35 | <wanderview> | but some kind of signing on ntp would be cool too |
| 15:07 | <annevk> | Hmm, halfway rewriting file URL parsing and my Windows VM refuses to boot |
| 15:20 | <annevk> | So now I can't perform the final checks around Windows drive letter until I download a new 5GiB blob and make it run... |
| 15:20 | <annevk> | More tomorrow I guess... |
| 15:32 | <Domenic> | annevk: I'll run tests for you! |
| 15:32 | <Domenic> | Sebmaster: ^ movement on file URLs |
| 15:32 | <Domenic> | (in addition to the recent commits) |
| 15:39 | <annevk> | Domenic: put a list in https://www.w3.org/Bugs/Public/show_bug.cgi?id=23550#c3 |
| 15:40 | <annevk> | Domenic: but I think I'll call it a day for now and take a break from file URLs until tomorrow |
| 15:41 | <annevk> | Domenic: my head is feeling rather fried |
| 15:41 | <Domenic> | annevk: ok, sounds good! |
| 15:42 | <Sebmaster> | Yeah i saw on twitter |
| 16:59 | <annevk> | TabAtkins: url.bs has some hardcoded references to Encoding through class=anchors |
| 16:59 | <annevk> | TabAtkins: some of those "conflict" with DOM Standard terms |
| 16:59 | <annevk> | TabAtkins: resulting in warnings |
| 16:59 | <annevk> | TabAtkins: how can I avoid the warnings? |
| 17:00 | <TabAtkins> | They'll only conflict if they're not distinguishable (same text/type/for) |
| 17:00 | <TabAtkins> | Are you doing that on purpose? |
| 17:01 | <annevk> | TabAtkins: Encoding is not yet in Bikeshed |
| 17:01 | <annevk> | TabAtkins: not sure why DOM does not use type/for |
| 17:01 | <annevk> | TabAtkins: I think it does actually |
| 17:02 | <TabAtkins> | I'll look at it, shortcut this process. ^_^ |
| 17:02 | <annevk> | Yeah, DOM uses for=Document |
| 17:07 | <TabAtkins> | annevk: Oh, it's not the definitions conflicting (directly), it's your autolinks. They're just <a>encoding</a>, which doesn't let Bikeshed know which of the multiple local "encoding" dfns it should pick. |
| 17:07 | TabAtkins | is considering moving anchors-block things into an intermediate category between local and remote, to avoid this problem happening accidentally. |
| 17:08 | <annevk> | TabAtkins: when I change to <a dfn for>encoding</a> nothing changes |
| 17:08 | <TabAtkins> | Yeah, empty for is no information. :/ Try <a for="/">encoding</a> |
| 17:08 | <annevk> | TabAtkins: but also, it seems that if you define a class=anchors block you should get the shortcut |
| 17:08 | <TabAtkins> | What shortcut? |
| 17:08 | <annevk> | TabAtkins: not having to define anything |
| 17:09 | <annevk> | class=anchors > external stuff |
| 17:09 | <TabAtkins> | Yes. |
| 17:09 | <TabAtkins> | Currently I do that by making them equivalent to local dfns. |
| 17:09 | <annevk> | TabAtkins: so if I need for=/ |
| 17:09 | <annevk> | TabAtkins: the warning I get is wrong |
| 17:09 | <annevk> | TabAtkins: it says "Arbitrarily chose the one with type 'dfn' and for ''" |
| 17:09 | <annevk> | TabAtkins: which suggests <a for> to me |
| 17:10 | <TabAtkins> | Ah, indeed. You're totally right about that. That's on me, I'll fix the warning. |
| 17:13 | <TabAtkins> | Hm, actually, I'm only finding one local encoding ref in url.bs - the one from Encoding. I'll need to look into this more, don't make any changes yet. |
| 17:14 | <annevk> | Yeah, using for=/ doesn't help |
| 17:15 | <annevk> | If I remove the ones from class=anchors though I do get concept-document-encoding from the DOM Standard |
| 17:15 | <annevk> | Weird... |
| 17:16 | <TabAtkins> | Yeah, bug on my part - somehow I'm inserting *two* versions of the Encoding ref into the local anchors db. They're identical, so dunno what's happening, but that's why Bikeshed is complaining about finding multiple of them. ^_^ |
| 17:17 | <TabAtkins> | UGH I FOUND IT |
| 17:17 | <TabAtkins> | "encode" and "encoding" are both returned by the fuzzy matcher for the autolink "encoding". |
| 17:18 | <TabAtkins> | Hm, I wonder how to fix this. |
| 17:25 | <TabAtkins> | All right, I know how to fix. No change required on your part. |
| 22:45 | <ChALkeR> | Hi all. |
| 22:45 | <ChALkeR> | I suppose that there is a mistype in the css3-values spec. |
| 22:46 | <ChALkeR> | Regarding the 1px size. |
| 22:48 | <ChALkeR> | http://www.w3.org/TR/css3-values/#absolute-lengths |
| 22:49 | <ChALkeR> | > The image below illustrates the effect of viewing distance on the size of a reference pixel: a reading distance of 71 cm (28 inches) results in a reference pixel of 0.26 mm, while a reading distance of 3.5 m (12 feet) results in a reference pixel of 1.3 mm. |
| 22:49 | <ChALkeR> | That does not match with the numbers in the image. |
| 22:49 | <ChALkeR> | http://www.w3.org/TR/css3-values/pixel1.png |
| 22:50 | <TabAtkins> | All of them match except it has 138 inch rather than 144 |
| 22:50 | <TabAtkins> | Which is 11.5 feet, close. |
| 22:50 | <TabAtkins> | (And is the actual closest inch measurement to 3.5m) |
| 22:50 | <ChALkeR> | 0.28mm and 0.26mm |
| 22:50 | <TabAtkins> | Ah, maybe the old /TR version is different but with newer text. |
| 22:51 | <ChALkeR> | 1.4mm and 1.3mm |
| 22:51 | <TabAtkins> | Don't ever read /TR anyway. https://drafts.csswg.org/css-values/#absolute-lengths |
| 22:51 | <ChALkeR> | Ah, that one has a fixed image, thanks =) |
| 22:51 | <TabAtkins> | np |
| 22:53 | <ChALkeR> | Looks like TR has an old image, compare http://www.w3.org/TR/css3-values/pixel1.png and https://drafts.csswg.org/css-values/pixel1.png |
| 22:53 | <TabAtkins> | Yeah, the TR one wasn't updated in time I guess. |
| 22:53 | <TabAtkins> | It'll be fixed next time we publish. |
| 22:53 | <ChALkeR> | Ok, thanks =)/ |
| 22:54 | <ChALkeR> | («/» is there by an accident) |
| 23:25 | <Domenic> | ChALkeR: TR stands for trash, never read anything with that in the URL |
| 23:33 | <ChALkeR> | lol :-) |
| 23:36 | <boogyman> | Domenic: may I suggestion to add a "dont index" header to things rendering on those domains. Google search: "css spec" gives the CSS 2.1 (TR) link as the first result |
| 23:36 | <Domenic> | boogyman: believe me, if I controlled those domains I would. |
| 23:37 | <boogyman> | at least it's w3.org and not the schools variant |
| 23:44 | <ChALkeR> | boogyman: Only in the case if you don't omit the «spec» part… |
